any word on the ibike unit?

Discussion in 'Power Training' started by mmerchant, May 17, 2006.

  1. ahaile

    ahaile New Member

    Joined:
    Aug 6, 2003
    Messages:
    63
    Likes Received:
    0
    That was me on the iBike list. Did nhorowitz send you a WKO file or the iBike app's .csv file? I believe that WKO converts all power values back to integers. If it's the .csv file, scroll through it and you'll see some non-integer sections that are from interpolated values. I haven't seen a power file yet with no dropouts. This won't catch all interpolated sections, however, since if the gap being spanned happens to break down into integer increments (e.g., a 15 watt jump over two missed records, ramping linearly as +5 and +10) then you'll see integers there too. When I was using this method to quantify the total amount of dropouts, I had to make a calculation to figure out how many dropouts I probably missed based on the histogram of dropouts I found.

    I've noticed the correlation b/w wild power numbers and tilt too.
     


  2. acoggan

    acoggan Member

    Joined:
    Jul 4, 2003
    Messages:
    3,047
    Likes Received:
    9
    Sorry...clearly, I participate in too many lists if I can't even keep the other posters straight!

    Anyway, the files I received were in the .wko format, and I believe that they were created by downloading via the iBike software and then imported into WKO+. Based on what you say, it sounds like the section of the file that's causing the problem may in fact be the result of interpolation, as it spans a number of records during which power is reported as increasing quite linearly.
     
  3. ahaile

    ahaile New Member

    Joined:
    Aug 6, 2003
    Messages:
    63
    Likes Received:
    0
    No problem. I don't think ANYONE is going to complain about you participating in too many forums, Andy!

    That's been the case in similar files I've looked at. You might be interested in two files I posted on the iBike list a few days ago. There's one segment (labeled "Bad 5" in the .wko files) where I experienced a similar problem, though not as severe as what nhorowitz reported. One file uses linear interpolation across the dropouts, while the other zeroes them. Neither, of course, could be called "right" -- what is "right" when we're dealing with no and/or bad data? -- but they demonstrate the effect different interpretations of dropouts have on AP and NP.

    "Bridged" dropouts: http://www.duke.edu/~aah3/iBike-bridged.wko
    Bad 5: 360AP / 446 NP
    Entire ride: 234AP / 276 NP

    "Zeroed" dropouts: http://www.duke.edu/~aah3/iBike-zeroed.wko
    Bad 5: 198AP / 228 NP
    Entire ride: 219AP / 253 NP

    I'm not sure what the best way to deal with dropouts is. Presumably, whatever method is most neutral to AP and NP. Maybe replace the missing segments with the AP value? Or suture together the existing data and base AP and NP on that? That last sounds like far too much trouble though, like good money after bad.
     
  4. RChung

    RChung New Member

    Joined:
    Sep 12, 2006
    Messages:
    306
    Likes Received:
    1
    The best way to deal with dropped out data is to mark as "missing data", and not to interpolate or set to an arbitrary value (like zero).
     
  5. ahaile

    ahaile New Member

    Joined:
    Aug 6, 2003
    Messages:
    63
    Likes Received:
    0
    Interesting. How would you then handle the 30-second rolling average of the NP algorithm? Do you have to just say that NP doesn't exist for 30 seconds either side of the gap?
     
  6. RChung

    RChung New Member

    Joined:
    Sep 12, 2006
    Messages:
    306
    Likes Received:
    1
    Imputation of missing values is not a canned approach. It depends on what you need the data for, how smooth the power was on either side of the gap, and whether there was any other info (like, for example, cadence, speed, elevation, or wind speed) that you can leverage off of. But you can't do any of that stuff unless you know exactly where the missing data are, and interpolating or filling in with zeroes doesn't let you do that.
     
  7. ahaile

    ahaile New Member

    Joined:
    Aug 6, 2003
    Messages:
    63
    Likes Received:
    0
    I'm just reviving this thread to mention that the newly released iBike firmware, 1.08, does exactly what you were mentioning, Robert. When the vibrations overload the accelerometer, rather than give up on calculating power it goes into a fallback mode. Instead of getting slope and acceleration data from the accelerometer, it uses the last known good reading for slope and calculates acceleration from changes in velocity read at the front wheel sensor. So long as slope doesn't change before the vibrations settle down, this should be pretty accurate. If it happens at a point when slope is changing -- like a fast rolling chipseal downhill -- then you'll see errors, sometimes high sometimes low depending on whether slope is falling or rising after the vibrations start.
     
  8. stea1thviper

    stea1thviper New Member

    Joined:
    Nov 11, 2006
    Messages:
    11
    Likes Received:
    0
    Hey everyone,

    So I've been reading up on various forums around the internet for days now on this ibike powermeter. I've still got a couple questions though that I could not find answers to. Given the fact that I am a poor college student who trains seriously but wants to also train effectively, the ibike seems like the way to go. I am concerned however as to it's accuracy in pack where there is alot of turbulance the ibike will not be able to pick up. Does anyone have experience with this?

    I'm also concerned with the varying power outputs with different riding positions. I tend to go from the hoods to the tops to out of the saddle very often throughout my rides, and sometimes into the drops as well with strong headwinds. Essentially I have no one "main" riding position. Any idea how badly this would scewer data?

    Money is a pretty tight issue for me, so I've pretty much narrowed it down to this ibike, or an older used powertap, which I could get for around the same price. I've heard alot of problems with water and such with the older powertaps however so I'm still pretty undecided.

    -Steve
     
  9. RChung

    RChung New Member

    Joined:
    Sep 12, 2006
    Messages:
    306
    Likes Received:
    1
    Interesting. Does it mark the imputed values as having been imputed?
     
  10. ahaile

    ahaile New Member

    Joined:
    Aug 6, 2003
    Messages:
    63
    Likes Received:
    0
    Not that I've been able to tell.
     
  11. ahaile

    ahaile New Member

    Joined:
    Aug 6, 2003
    Messages:
    63
    Likes Received:
    0
    Hi Steve -- A lot seems to have changed with the iBike's new firmware, v1.08, so the answers to some of your questions may be in flux. I'll see if I can offer anything useful from my experience with it.

    Turbulence doesn't seem to be the problem. The problem the iBike has with drafting is that, when you're only behind 1 or 2 riders, the draft where the iBike sits (right behind the lead rider's biggest area) is better than it is at the extremities, so the iBike overestimates how much of a draft you're getting. In a pack, there are enough riders around you that the draft is everywhere and the error is smaller. That's the hypothesis, anyway. Many of us users have noticed that watts seem low when in a small paceline but seem better in a pack, but no one, that I know, has done a side-by-side comparison to another powermeter to figure out just how much the difference is.

    I've only had one group ride on the new firmware, and, subjectively, it seems like the iBike is doing a lot better with drafting. They've definitely improved the reported windspeed with this release, and some of those improvements may be spilling over to the drafting issue.
    What I did was calibrate my iBike while riding in the drops. My reasoning is that the faster you go, the more of your power is going into air drag, so I want my unit to be accurate on air drag in that situation. Since I'm usually in the drops when I'm going fast, that's how I did my coastdown. As you slow down, air drag becomes a smaller component of the total, so an error there produces a smaller error to the total power calculation.
    On the budget end of powermeters, Polar is coming out with an updated cyclecomputer and power sensor that also look interesting. MSRP is $650 and availability is supposed to be May.
     
  12. genedoc

    genedoc New Member

    Joined:
    Aug 15, 2006
    Messages:
    63
    Likes Received:
    0
    And don't forget that Performance sells the PowerTap WITH a built wheel for $1000. Wait for one of their fairly frequent 20% off deals and join their Team Performance deal for 10% back in store credit and you can get a new PowerTap with wheel for $720. At that price, purchasing an iBike ($500 with cadence and no wheel) or the Polar ($650 with no wheel) becomes more a matter of device preference and less of a "bargain".

    I'm still trying to decide, but the only downside I can see to the PT is it's fixed to one wheel.
     
  13. ahaile

    ahaile New Member

    Joined:
    Aug 6, 2003
    Messages:
    63
    Likes Received:
    0
    Well, to compare apples to apples. those same 20% and 10% discounts apply to the iBike and Polar too, so it's $720 for the Powertap vs. $288 for the iBike ($378 w/ cadence) and $478 for the Polar. But I agree that, within that range, you should pick the one you want most.
     
  14. stea1thviper

    stea1thviper New Member

    Joined:
    Nov 11, 2006
    Messages:
    11
    Likes Received:
    0
    wow thanks for the excellent replies. i can definitely wait until performance has another 20% off sale. polar's new power device sounds interesting, although i couldnt find any extra info on it. any guesses as to how it measures power? (hopefully not the chain vibration again)
     
  15. genedoc

    genedoc New Member

    Joined:
    Aug 15, 2006
    Messages:
    63
    Likes Received:
    0
    If Performance sold those units.
     
  16. ahaile

    ahaile New Member

    Joined:
    Aug 6, 2003
    Messages:
    63
    Likes Received:
    0
  17. genedoc

    genedoc New Member

    Joined:
    Aug 15, 2006
    Messages:
    63
    Likes Received:
    0
  18. VeloPower

    VeloPower New Member

    Joined:
    Dec 1, 2006
    Messages:
    4
    Likes Received:
    0
    iBike Pro Review.
    I have an iBike and a Powertap and I have been comparing them side by side and here are the conclusions:

    The iBike is NOT useful for training it is strictly a gadget and here is why:
    The iBike is not at all accurate real time. Look at the graphs above and you will see that MOST OF THE TIME the iBike is significantly (often 50-100 Watts) high or low. If you have ever (seriously) trained to power then you know that just a few watts (let's say 10% of your max hour of power; 25-50 Watts unless you are Lance Armstrong prior to 2006 (sorry Lance)) high or low makes a HUGE difference when you are trying to hold a set wattage for a while; and with this much error, you will have a horrible workout with the iBike. Sure, the marketing folks at iBike like to point to average watts and say that it is accurate but this is not relevant to real time training on the bike. The iBike is USELESS as a real time wattage training tool. There are also several other serious problems for example:
    1) What about WINTER time ??? YOU CAN'T USE THE iBIKE INDOORS ! You will have to buy 2 power meters anyhow so what good is it ?
    2) The iBike is completely inaccurate on the slightest rough road.
    3) Turns. The iBike is flat out wrong when going through turns.
    4) Drafting. Yes, the marketing team at iBike want you to think that it is more accurate when in bigger packs. BS ! I have tried it. It is WRONG WRONG WRONG in the draft.

    Conclusion, If:
    1) You just like gadgets and don't care if they really work.
    2) You never turn.
    3) You never ride in the draft.
    4) You don't mind a significantly wrong watts readout while you are training.
    5) You are a sucker and believe all of the BS that iBike and those that stand to benefit from iBike sales tell you.
    Then go buy an iBike -- There are plenty of them for sale on eBay from the people who have gotten suckered into buying one and can't get a refund.
    Otherwise, if you are smart, go buy a PowerTap, Ergomo, SRM etc...
     
  19. jhamann

    jhamann New Member

    Joined:
    Dec 2, 2006
    Messages:
    2
    Likes Received:
    0

    Sorry to hear you're not happy with your iBike. I suppose there isn't a product anywhere that pleases 100% of the people 100% of the time.

    Fortunately, there are many, many, many other iBike owners, who also own other power meters, who also have done side-by-side testing, and who have come to exactly the opposite conclusions you have reached. Comparative data is posted on our website https://www.ibikesports.com/documents/Which_is_the_iBike.pdf for all to see.

    You state that the iBike is completely inaccurate on the slightest rough road. That can mean only one thing: you haven't installed the free firmware update that fixes the rough road problem and improves the watts accuracy of the iBike. You should give it a try; I think many of your concerns would be diminished.

    Yet, even with this latest firmware you're certainly correct that the iBike has limitations:

    1) At present the iBike doesn't work on an indoor trainer; we've never said otherwise. It's odd that you criticize the iBike for this limitation yet you own one yourself.

    2) We state quite explicitly in the instruction manual that the iBike isn't accurate in sharp turns (such as turning at street intersections). If it's critical for you to know with highest accuracy how many watts you're expending while making a left or right turn then the iBike isn't for you.

    3) Drafting: there are forum postings that report good accuracy in large packs and lesser accuracy in smaller packs. Do you have additional data? If so please share the data, not just the conclusion.

    I checked ebay this morning. There is one man selling his iBike (I know him, he is a very nice guy) and there is one online retailer. Doesn't sound to me as if the masses have decided they've been suckered...

    Refunds? If you're this unhappy return your iBike to us and we'll give you a refund. Out of the thousands of units we've sold only two people have asked for refunds. You can be number three.

    John Hamann
     
  20. VeloPower

    VeloPower New Member

    Joined:
    Dec 1, 2006
    Messages:
    4
    Likes Received:
    0
    Well hello Mr. Hamann. It appears that you wish to put me into the category of an unhappy customer so that it will diminish the value of my review. I assure you that my only goal was to put out an objective review on your product. I don't require a refund; thanks anyway. You have done a great job of spin doctoring and sugar coating all of the significant problems with your product and I felt that it was time that someone did an objective unbiased review.
    1) The graphs on the link you gave:
    https://www.ibikesports.com/documents/Which_is_the_iBike.pdf
    all make the point very well about how the iBike is constently high or low a significant amount while riding the bicycle and thus make it useless as a real time training tool. Thank you for helping to make that point. In addition to the graphs it is even more of a glaring problem when actually riding the bicycle with the two power meters side by side and trying to hold a constant wattage that it becomes very apparent that the iBike drifts high and low making it worthless for serious training.
    2) Turns. You are trying to SPIN DOCTOR turns to sound better than they really are by saying that it doesn't work in SHARP turns. Well that's BS. During my testing, I found that the iBike was wrong in pretty much every turn on my training course, most of which I would not consider to be SHARP.
    3) Drafting. You are trying to SPIN DOCTOR that too. I tried the iBike in numerous drafting conditions and it is just plain wrong. My guess is that those posting in the forums that you refer to (if they even exist) did not have two power meters side by side for comparison. I did and the fact is that the iBike is not accurate in the draft.
    4) Ebay. Go to www.ebay.com. Search for iBike Pro. Make sure and check the box entitled "Completed Listings". From Nov 19 to Nov 30 (11 days) there have been 6 iBikes for sale. That's one every two days on average. Doesn't look too good for a product with small volumes.

    In conclusion, my review is based on putting the iBike and PT side by side in real life training, racing riding situations. The folks at iBike are great at smoke and mirrors and spin doctoring and they want people to think their product is much better than it really is. I hope this review helps to debunk some of the iBike myths. So if you want a gadget, then go buy an iBike -- there are probably some people who are happy with the iBike as a gadget. If you are a serious athlete, get something else...
     
Loading...
Loading...